|
Post by venkat3271 on Oct 18, 2007 5:34:49 GMT -5
Hi everybody, By nyquist Sampling theorem, Fs=2 * fm.... But when i sample a sine at t=0 with the above condition i cant get back the signal... What makes this limitation?..
|
|
|
Post by isomorphism on Dec 22, 2007 6:55:25 GMT -5
Probably sampling theorem could be applied to signals with no spectral impulses. Meaning it should have a non-zero spectral width. Almost all non-harmonic signal do so. Moreover in my opinion, if you consider choosing a random starting point of sampling at a uniform rate, you can almost all the time reconstruct with reasonable accuracy. As usual with samples, it's more the merrier
|
|
|
Post by mohasin on Feb 21, 2008 2:26:36 GMT -5
i think its because when u sample a sine at t=0 with fs=2*fm samling ocuurs at the zero crossing of the signal. so sample it at fs> 2*fm
|
|
|
Post by prasadrayi on Feb 22, 2008 3:14:40 GMT -5
the sampling time period should be Ts< 1/2fm , so ampling at zero crossigs of the signal
|
|
micky
Junior Member
Posts: 5
|
Post by micky on Apr 17, 2008 8:28:52 GMT -5
According to the Nyquist sampling theorem (which is easy to prove), if you sample a signal at a rate which is no less than TWICE its highest frequency component, then you can reconstruct it with no loss of information. This is true.
Your example of sampling a single sinewave at the exact instances in time when it crosses the time axis and has a zero value is an extreme case and, of course, you will then reconstruct absolutely nothing. This is not a shortcoming of the sampling theorem but a footnote to it that says "don't sample at only the exact instances when it crosses the time axis".
Michael
|
|
|
Post by charan langton on May 10, 2008 11:23:03 GMT -5
The Nyquist Theorem says that the sampling speed must be greater than two times the largest frequency, which gaurantees that you will not get two zero crossings. Charan
|
|
|
Post by abhityagi85 on Jul 6, 2008 10:38:54 GMT -5
ur question is very genuine. but actually as u said u r sampling at 2*fs rate. and sampling from instant t=0. so here u are sampling your signal when the value of signal is 0. it is equivalent to no signal for sampler. isnt it? thus whatever samples u get after sampling (0 in this case) are simply reconstructed but it contradicts to your input signal. so the systems employ sampling at certain higher rates to avoid sampling at 0 signal instant.
|
|